elib
DLR-Header
DLR-Logo -> http://www.dlr.de
DLR Portal Home | Imprint | Privacy Policy | Contact | Deutsch
Fontsize: [-] Text [+]

Non-Parametric Calibration for Classification

Wenger, Jonathan and Kjellström, Hedvig and Triebel, Rudolph (2020) Non-Parametric Calibration for Classification. In: 23rd International Conference on Artificial Intelligence and Statistics, AISTATS 2020. International Conference on Artificial Intelligence and Statistics (AISTATS), 26. - 28. August 2020, Virtual.

[img] PDF
1MB

Abstract

Many applications of classification methods not only require high accuracy but also reliable estimation of predictive uncertainty. However, while many current classification frameworks, in particular deep neural networks, achieve high accuracy, they tend to incorrectly estimate uncertainty. In this paper, we propose a method that adjusts the confidence estimates of a general classifier such that they approach the probability of classifying correctly. In contrast to existing approaches, our calibration method employs a non-parametric representation using a latent Gaussian process, and is specifically designed for multi-class classification. It can be applied to any classifier that outputs confidence estimates and is not limited to neural networks. We also provide a theoretical analysis regarding the over- and underconfidence of a classifier and its relationship to calibration, as well as an empirical outlook for calibrated active learning. In experiments we show the universally strong performance of our method across different classifiers and benchmark data sets, in particular for state-of-the art neural network architectures.

Item URL in elib:https://elib.dlr.de/135322/
Document Type:Conference or Workshop Item (Speech)
Title:Non-Parametric Calibration for Classification
Authors:
AuthorsInstitution or Email of AuthorsAuthor's ORCID iD
Wenger, JonathanUniversity of Tübingenhttps://orcid.org/0000-0003-2261-1331
Kjellström, HedvigKTH Royal Institute of Technologyhttps://orcid.org/0000-0002-5750-9655
Triebel, RudolphRudolph.Triebel (at) dlr.dehttps://orcid.org/0000-0002-7975-036X
Date:August 2020
Journal or Publication Title:23rd International Conference on Artificial Intelligence and Statistics, AISTATS 2020
Refereed publication:Yes
Open Access:Yes
Gold Open Access:No
In SCOPUS:No
In ISI Web of Science:No
Status:Published
Keywords:Supervised deep learning; Classification; Uncertainty Estimation; Gaussian Processes
Event Title:International Conference on Artificial Intelligence and Statistics (AISTATS)
Event Location:Virtual
Event Type:international Conference
Event Dates:26. - 28. August 2020
HGF - Research field:Aeronautics, Space and Transport
HGF - Program:Space
HGF - Program Themes:Space System Technology
DLR - Research area:Raumfahrt
DLR - Program:R SY - Space System Technology
DLR - Research theme (Project):R - Vorhaben Multisensorielle Weltmodellierung (old)
Location: Oberpfaffenhofen
Institutes and Institutions:Institute of Robotics and Mechatronics (since 2013) > Perception and Cognition
Deposited By: Triebel, Rudolph
Deposited On:25 Nov 2020 09:44
Last Modified:25 Nov 2020 09:44

Repository Staff Only: item control page

Browse
Search
Help & Contact
Information
electronic library is running on EPrints 3.3.12
Copyright © 2008-2017 German Aerospace Center (DLR). All rights reserved.